Anorexia nervosa (AN) and bulimia nervosa (BN) are prevalent disorders in adolescent and young adult women and are associated with high rates of morbidity and mortality. There are no other prospective, long-term studies with interviews conducted at frequent intervals which map the course and outcome of these disorders. We propose to extend our current five-year, prospective, naturalistic, longitudinal study of 250 women with AN and/or BN for up to 11 years. The aims are: 1) to describe longitudinal patterns of the course of each disorder; 2) to ascertain whether AN and BN are distinct phenomena; 3) to assess prognostic factors; and 4) to assess the relationship between eating disorders and comorbid disorders. In the first five years of the study, we recruited 250 subjects with AN, BN, or a combination of the two (AN/BN), completed the schedule of interviews described in the original proposal, assembled master files of intake and three, six, nine, 12, 15, 18, and 21 month follow-up data, analyzed intake, follow-up, and treatment data collected to date, designed and completed four independently funded pilot studies on our cohort, and conducted one reliability study. The attrition rate was six percent. The sample was characterized by high levels of current comorbid Axis I disorders (71%) and relatively low rates of Axis II (27%) disorders. While comorbidity did not differentiate AN and BN, course of illness did. In the first year and a half of follow-up, a majority of subjects improved symptomatically. Rates of full recovery (asymptomatic for greater than or equal to 8 weeks) were: 7.6% of AN, 46.4% of BN, and 18.4% of AN/BN subjects. Rates of partial recovery (less than full DSM-III-R criteria for greater than or equal to 8 weeks) were: 51.6% of AN, 79.3% of BN, and 61.1% of AN/BN (mixed AN and BN) subjects. Eating disorder type and percent ideal body weight emerged as significant predictors of partial recovery, while comorbidity and duration of the intake eating disorder episode did not. We observed high rates of psychotherapy and, particularly in the persistently ill bulimics, relatively low rates of pharmacotherapy. The data suggest that the diagnosis of AN has serious implications. Continued longitudinal follow-up is essential in order for a sufficient number of subjects to recover, which will thereby allow the accurate assessment of relapse, recurrence, and predictors. We are proposing to continue conducting quarterly evaluations for an additional five years on all probands using semi-structured diagnostic interviews. The proposed study will assist in treatment planning, improve our design of research protocols on pathophysiology and treatment, increase patient education about AN and BN, and contribute to nosological classification of the eating disorders.